Planning Using Combinatory Categorial Grammars

This paper presents a new model of planning based on representing domain knowledge using Combinatorial Categorial Grammars taken from natural language processing. This enables the capturing of plans with context-free expressiveness. It uses the same representation that has previously been used for plan recognition and has been shown to be learnable. Thus it represents a solid link between planning, plan recognition, and natural language processing. The paper also compares our open source implementation to two other well known hierar-